ABSTRACT
Although the structure of the catalytic RNA component of ribonuclease P has been
well characterized in Bacteria, it has been little studied in other organisms,
such as the Archaea. We have determined the sequences encoding RNase P RNA in
eight euryarchaeal species:
Halococcus morrhuae
,
Natronobacterium gregoryi
,
Halobacterium cutirubrum
,
Halobacterium trapanicum
,
Methanobacterium thermoautotrophicum
strains
[Delta]
H and Marburg,
Methanothermus fervidus
and
Thermococcus celer
strain AL-1. On the basis of these and previously available sequences from
Sulfolobus acidocaldarius
,
Haloferax volcanii
and
Methanosarcina barkeri
the secondary structure of RNase P RNA in Archaea has been analyzed by
phylogenetic comparative analysis. The archaeal RNAs are similar in both
primary and secondary structure to bacterial RNase P RNAs, but unlike their
bacterial counterparts these archaeal RNase P RNAs are not by themselves
catalytically proficient
in vitro
.
Ribonuclease P (RNase P) is an endoribonuclease present in all cells that
cleaves the leader sequences from precursor tRNAs to generate their mature 5'-ends (for reviews see
1
-
3
). RNase P has been most well studied in Bacteria, in which an RNA of ~130 kDa (400 nt) and a protein of ~14 kDa (120 amino acids) make up the holoenzyme (
4
). The RNA is the catalytic moiety and enzymatic activity by the RNA in the
absence of protein has been demonstrated
in vitro
from a wide phylogenetic range of Bacteria (
5
-
7
). RNase P from Archaea and Eucarya also contain an RNA component, but unlike
their bacterial counterparts, enzymatic activity by these RNase P RNAs when
separated from the other components of the holoenzyme or when synthesized
in vitro
has not been demonstrated (
8
,
9
).
Previous studies have shown significant biochemical diversity among the archaeal
RNase P enzymes. RNase P from the halophilic Archaea
Haloferax volcanii
is sensitive to micrococcal nuclease and has a high buoyant density (1.61 g/cm
3
) in cesium sulfate gradients (
10
). The halophilic RNase P appears, therefore, to be composed largely of RNA and
thus resembles the bacterial enzyme. RNase P from
Sulfolobus acidocaldarius
(misclassified as
S.solfataricus
in Darr
et al
.;
11
) is not sensitive to micrococcal nuclease, has a low buoyant density (1.27 g/cm
3
) in cesium sulfate and is much larger than the size of the RNA alone (
9
). RNase P from
S.acidocaldarius
is, therefore, predominantly protein and thus more closely resembles the
eucaryal RNase P. The genes encoding the RNAs associated with RNase P RNA
enzymes in
H.volcanii
(
8
),
S.acidocaldarius
(
9
) and
Methanosarcina barkeri
(D.W.Armbruster and C.J.Daniels, unpublished data, Genbank accession no.
U42984) have been cloned and characterized. All three contain recognizable
sequence similarity to one another and to bacterial RNase P RNAs, but not to
the eucaryal RNAs. Nevertheless, these archaeal RNase P RNAs are not capable of
catalysis in the absence of other components (presumed to be protein) of the
holoenzyme.
Secondary structure models of the archaeal RNase P RNAs presented previously
represent little more than the archaeal sequences `forced' into the model
developed for bacterial RNase P RNA secondary structure. In order to develop a
secondary structure model based on archaeal sequences we have cloned the genes
encoding RNase P RNA from eight additional species (see Fig.
1
):
Halococcus morrhuae
,
Natronobacterium gregoryi
,
Halobacterium cutirubrum
,
Halobacterium trapanicum
,
Methanobacterium thermoautotrophicum
strain [Delta]H,
M. thermoautotrophicum
strain Marburg,
Methanothermus fervidus
and
Thermococcus celer
strain AL-1 and used these sequences in a phylogenetic comparative analysis of
secondary structure.
Organisms were grown in the following media:
N.gregoryi
, natronobacteria medium (ATCC medium 1590);
H.morrhuae
, halobacterium medium (ATCC medium 1270);
H.cutirubrum
and
H.trapanicum
, yeast medium (ATCC medium 112) containing 25% (w/v) NaCl;
M.thermoautotrophicum
strains Marburg and [Delta]H and
M.fervidus
, ER medium (
12
). All cultures of extreme halophiles were grown aerobically at 37oC with continuous shaking; methanogens were grown without shaking under 40
p.s.i. 60% H
2
/40% CO
2
at 65 (
M.thermoautotrophicum
) or 80oC (
M.fervidus
). DNAs were purified as described previously (
13
). DNA from
T.celer
strain AL-1 was a gift from A.Reysenbach (Rutgers University).
Polymerase chain reactions (
14
) were performed using buffer containing 50 mM KCl, 10 mM Tris-HCl, pH 8.3, 1.5 [mu]M each dGTP, dCTP, dATP and dTTP, 0.05% Nonident P40, 5% acetamide
and 200 ng each primer oligonucleotide. The amplifications included an initial
2 min 94oC incubation and 30-40 amplification cycles (92oC for 1.5 min, 50oC for 1.5 min, 72oC for 0.5 min each cycle).
The genes encoding RNase P RNA from
M.thermoautotrophicum
strains Marburg and [Delta]H,
M.fervidus
and
T.celer
were obtained by amplification using primers A59FXba (GCTCTAGAGGAAAGTCCMSCC)
and A347RBam (CGGGATCCTAAGCCMSCTTYTGT). The resulting PCR products were
digested with
Xba
I and
Bam
HI and separated by electrophoresis in 3% low melting agarose gels (NuSieve GTG
agarose; FMC, Rockland, ME). Agarose plugs containing the DNA bands were
excised, melted (65oC) and used directly in ligation reactions containing restriction
endonuclease-digested pBluescript KS
+
DNA (Stratagene).
The gene encoding RNase P RNA in
H.trapanicum
and templates for enzyme assays (described below) were obtained by
amplification using primers T7 P-PCR 5' (TAATACGACTCACTATAGGCAGAGAGAGCCCGGC) and P-PCR 3' (AAMRYGGCTGAGAGGAGTAAGCC). The PCR products were
purified as above and cloned into pUC19 (BRL).
RNase P RNA genes from the other halophilic species were cloned intact from
genomic DNA using methods described previously (
13
). The RNA transcript used as probe to identify DNA fragments containing RNase P
RNA-encoding genes was generated from a
Mae
I/
Bst
B1 subclone of the
H.volcanii
RNase P RNA gene-containing plasmid pIBI31::MB (
8
) using T7 RNA polymerase. Restriction endonuclease digests of genomic DNAs were
separated by agarose gel electrophoresis and fractions of the gels, previously
shown by Southern analysis to contain putative RNase P RNA-encoding genes, were excised. DNA was recovered from the gel slices either
by disruption of the gel with NaI and purification of the DNA with glass powder
(
15
) or by dilution and phenol extraction from low gelling temperature agarose. The size-selected DNAs were then cloned into plasmid vectors, transformed into
Escherichia coli
DH5[alpha]F' and screened by colony hybridization with the same probes used in
Southern analyses. Nucleotide sequences of the cloned DNA fragments and
Northern and Southern hybridizations were used to verify that the cloned DNAs
encoded RNase P RNAs from each organism.
The nucleotide sequences encoding RNase P RNAs were determined from double-stranded plasmid DNAs by the dideoxy chain termination method (
16
) with Sequenase version 2.0 (Amersham, Arlington Heights, IL) using M13
universal, M13 reverse, A59FXba and A347RBam primers and in some cases
exonuclease III-generated nested deletions (
17
). 7-Deaza-dGTP was used to alleviate band compressions in sequencing gels.
The activity of RNase P RNA alone enzyme function was measured in enzyme assays
performed at 37oC for 1 and 24 h in the presence 25 mM MgCl
2
, 50 mM Tris-HCl, pH 8, 0.1% SDS, 0.05% Nonident P-40, 250 mM-3 M NH
4
Cl, 20 nM RNase P RNA and 20 nM precursor tRNA (
18
).
Escherichia coli
RNase P RNA was synthesized using plasmid pDW98 (
18
) for use in control reactions. Internally
32
P-labeled precursor tRNA
Asp
from
Bacillus subtilis
was synthesized from plasmid pDW128 (
18
).
Templates for the synthesis by T7 RNA polymerase of halophile RNase P RNAs were
PCR amplification products from plasmid DNAs. Amplification was from pIBI31::MB
(
H.volcanii
), pB-171 (
H.morrhuae
), p-28 (
N.gregoryi
), pRNA (
H.cutirubrum
) and pT7Trap (
H.trapanicum
) using primers T7 P-PCR 5' and P-PCR 3' (see above). PCR-generated templates were extracted with
phenol/chloroform and treated with the Klenow fragment of
E.coli
DNA polymerase I in the presence of excess dNTPs prior to
in vitro
transcription using T7 RNA polymerase.
Partial RNase P transcripts (~1 [mu]g) from the methanogens and
T.celer
were tested after reconstitution, by heating to 80oC and slow cooling, with 1 [mu]g of an oligonucleotide designed to anneal to up- and downstream polylinker-derived sequences and the downstream strand of P2 to
artificially generate secondary structure absent in the partial clones (see
Fig.
2
). The homologous
E.coli
partial RNA for use in control reactions was generated by transcription from a
clone containing the partial sequence amplified from
E.coli
DNA using 59FBam and 347REco (the bacterial versions of A59FXba and A347RBam) (
19
).
Sequences were aligned manually using SeqApp (Don Gilbert, Indiana University).
Comparative analysis of secondary structure was preformed as previously
described (
21
). Comparative support for secondary structure was identified using Covariation
(
22
). Sequences derived from primers in clones obtained by PCR amplification were
excluded from the analysis. The structures of regions unique to each sequence
(P12, P15/P16 and P19) and therefore not amenable to comparative analysis were
predicted thermodynamically using Mulfold (
22
). All sequences, alignments and secondary structures are available from the
Ribonuclease P Database (http://jwbrown.mbio.ncsu.edu/RNaseP/). The Genbank
accession nos for these sequences are U42980, U42981, U42982, U42983, U42985,
U42986, U42987 and U42988.
A subclone of the
H.volcanii
RNase P RNA gene (
8
) was used as probe in Southern analyses to identify DNA fragments encoding
RNase P RNA in genomic DNAs of
H.cutirubrum
,
H.morrhuae
and
N.gregoryi
. The
H.volcanii
probe hybridized to single fragments in each of several restriction digests of
the genomic DNAs, indicating that RNase P RNA is encoded by a single copy gene
in each of these halophiles, as is also the case for previously cloned archaeal
genes and those of Bacteria (
4
). The DNAs containing the
H.morrhuae
(2.2 kb
Sal
I fragment ),
H.cutirubrum
(7.6 kb
Sph
I fragment) and
N.gregoryi
(2 kb
Xma
I fragment) RNase P RNA-encoding genes were cloned and the nucleotide sequences of the RNase P RNA-encoding regions contained within these fragments were determined.
The 5'- and 3'-terminal sequences of the
H.volcanii
,
H.morrhuae
,
H.cutirubrum
and
N.gregoryi
RNase P RNA-encoding genes are highly conserved and so these sequences were used to
clone the gene from
H.trapanicum
genomic DNA by PCR amplification. Each of these cloned genes hybridized to RNAs
of the expected sizes in Northern analyses and to the corresponding single DNA
fragments of the source genomic DNA (data not shown), confirming that the
cloned fragments encode functional RNase P RNAs. Weak hybridization in Northern
analyses to additional larger RNAs in the case of
H.trapanicum
(data not shown) suggests that this RNA may be synthesized as a larger
precursor molecule, as occurs in
E.coli
(
23
).
The RNase P RNA-encoding genes from
T.celer
,
M.thermoautotrophicum
strains Marburg and [Delta]H and
M.fervidus
were cloned by PCR amplification using primers complementary to highly
conserved sequences near the 5'- and 3'-termini of archaeal and bacterial RNAs. Analogous
primers have previously been used to obtain bacterial RNase P RNA-encoding sequences (
19
,
24
). In each case PCR resulted in the amplification of a single band in the
predicted size range (~300 bp). Each of the cloned genes hybridized to single DNA fragments of the
source genomic DNA (data not shown) in Southern analyses and phylogenetic trees
of the sequences are consistent with those based on 16S rRNAs (data not shown).
The archaeal sequences were analyzed by the comparative method (
20
) to construct a model for the secondary structure of archaeal RNase P RNA.
Compensatory changes were first identified amongst the sequences from the
extreme halophiles, which are sufficiently similar over the majority of their
lengths to be readily aligned on the basis of sequence similarity. This initial
secondary structure information was used to aid in addition of the methanogen
T.celer
and
S.acidocaldarius
sequences to the alignment. Several rounds of alignment/structure refinement
were performed, resulting in the construction of individual and consensus
secondary structures (Figs
2
and
3
). The consensus secondary structure is supported by compensatory changes at at
least two positions (the generally accepted indication that a helix exists) in
all helices except P5 and P10. P5 is supported by co-variation at a single base pair (U-A in
S.acidocaldarius
, A-U in all other cases). Evidence for or against the presence of P10 is
lacking; all four of the nucleotides in this helix are invariant in the
archaeal sequences available. Of 75 base pairings in the consensus secondary
structure 51 are individually supported by compensatory substitutions of the
paired nucleotides.
Although the secondary structure is for the most part well defined, the region
distal to P12 is extremely variable in length and sequence (even between close
relatives) and so no particular structure in this region is well supported.
These structures were therefore predicted thermodynamically (
22
), but it should be realized that no comparative evidence for these particular
structures is available. The region distal to P16 is also variable in sequence
length, although less dramatically than in the case of P12, and structure in
this region was likewise predicted thermodynamically, within the constraints of
the comparative data. The
M.fervidus
RNA lacks Watson-Crick complementarity in the nucleotides expected to form P6; this helix
is well supported amongst the remaining sequences, so this seems to be an
idiosyncrasy of this RNA.
The previously characterized RNase P RNAs from
H.volcanii
,
Methanosarcina barkeri
and
S.acidocaldarius
do not have the ability to cleave precursor tRNAs in the absence of other
enzyme components (
8
,
9
; D.W.Armbruster and C.J.Daniels, unpublished data, Genbank accession no. U42984). The products of the RNase P RNA-encoding genes from
H.cutirubrum
,
H.morrhuae
,
H.trapanicum
and
N.gregoryi
, synthesized
in
vitro
using T7 RNA polymerase, were tested for the ability to cleave precursor tRNA
Asp
from
B.subtilis.
Assays were performed at 37oC in standard assay buffer (
18
) at `low' (1 M ammonium acetate; optimal for the
E.coli
or
B.subtilis
RNAs) and `high' (3 M ammonium acetate; required by many mutant RNAs) ionic
strength. No specific cleavage of the substrate by the halophilic RNase P RNAs
could be detected, even upon extended incubation (24 h) and at an equimolar
RNase P RNA:substrate ratio (20 nM each) (data not shown).
The previous proposal for the secondary structure of an archaeal RNase P RNA,
that of
H.volcanii
, was constructed essentially by `forcing' the archaeal sequence into the model
for bacterial RNase P RNA structure; no other archaeal sequences were then
available for comparative analysis (
8
). In order to provide the information required to develop a model for RNase P
RNA secondary structure in Archaea we have determined the sequences of the
genes encoding RNase P RNA from eight members of the Euryarchaea: four species
of extreme halophiles, three species of methanogens and one sulfur-metabolizing thermophile. These sequences and those of
H.volcanii
,
S.acidocaldarius
and
M.barkeri
were analyzed by the comparative method (
20
) and a model for the secondary structure of archaeal RNase P RNA was developed
independently of the bacterial model.
In a comparative analysis of secondary structure potential helices in an RNA
molecule (i.e. complementary sequences) are tested by examination of the
equivalent nucleotides in homologous RNAs with different sequences. The
occurrence of sequence differences that nevertheless maintain complementarity
(i.e. compensatory changes) is evidence for the existence of a helix. The consensus archaeal RNase
P RNA secondary structure (Fig.
3
) is well supported by the comparative data. Of the 13 helices present in the
consensus model only two (P5 and P10) are not supported by compensatory changes
in two or more base pairings in the helix, the generally accepted indication
that a helix exists. In fact, of the proposed base pairings in the consensus
model 68% are specifically supported by compensatory sequence changes. There
are several regions of significant sequence length variation in the archaeal
RNase P RNAs; these correspond primarily to length variations in P3 and
structural variation in P12, P19 and the joining regions between P15/16 and P6.
Variation in these regions is not surprising, given that this is likewise seen
in the homologous regions in bacterial RNase P RNAs, however, the degree of
variation in P12 amongst the RNAs from the halophiles is striking given their
relatively close evolutionary affiliation.
The primary sequences and secondary structures of the archaeal RNase P RNAs
largely resemble their bacterial counterparts and those nucleotides which are
most conservative in the bacterial RNAs are also generally conserved in the
archaeal sequences. Despite these similarities, the archaeal RNAs are unlike
their bacterial counterparts in that they are not capable of catalysis in the
absence of other enzyme components. The structural basis of this deficiency
must reside in differences in the bacterial and archaeal sequences or higher
order structure. The archaeal consensus secondary structure contains all but
one helix present in the bacterial consensus structure, the 2 bp helix P11.
This structure is adjacent to a large region in which the structure remains
unknown even in bacterial RNase P RNAs. As in Bacteria, the presence of P19 is
sporadic, occurring only in the
M.barkeri
RNA amongst the archaeal RNAs.
The structure of the P15/16 region also varies amongst the archaeal RNAs (Fig.
2
), despite the conservation of this region in Bacteria, in which it has been
implicated in recognition of the substrate pre-tRNA 3'-terminus. In the halophile RNAs this region has the potential
to form a continuous helix of Watson-Crick (and G[middot]U) base pairs. However, in the methanogens and
T.celer
P15 and P16 are separated by an internal loop that conforms to the sequence and
structure consensus of the homologous region in the bacterial RNAs. Because of
its role in substrate recognition (at least in Bacteria;
25
,
26
), it is reasonable to imagine that structural divergence in this region, like
that seen in some of the Archaea, might be responsible for the absolute
dependence of these RNAs on the other (presumably protein) components of the
holoenzyme for catalytic activity. However, because even the methanogen RNase P
RNAs apparently lack catalytic proficiency
in vitro
, it seems unlikely that this region is entirely responsible for this `defect'
in the archaeal RNAs.
P18, a conservative element in bacterial RNase P RNAs, is entirely absent in the
archaeal structures. However, this element is also absent in the RNase P RNAs
of
Chlorobium tepidum
and
Chlorobium limicola
, two species of `green sulfur Bacteria'. The
Chlorobium
RNase P RNAs are nevertheless active in the absence of protein and deletion of
P18 from the
E.coli
RNA affects the optimal ionic strength, but not the kinetic properties, of that
RNA (
7
). Bacterial sequences lacking the usual 8 bp/GNRA loop form of P18 vary in P8
structure, which is otherwise highly conserved; this may, at least in part, be
attributed to a tertiary interaction between the P18 GNRA loop sequence and P8
(
24
). Archaeal P8 structure is also unusual, varying in both length and loop
structure, in accordance with the absence of P18.
There are also fine scale differences between conserved structures in the
archaeal and bacterial RNase P RNAs. P2, which is invariantly 7 bp in length in Bacteria (not including an unproven G[middot]U pair possible in all cases), is 6 bp in the Archaeal RNAs (in which
the additional pair is never possible). Perhaps associated with this change is
the difference in the length of the P2/P3 joining region (J2/3), which is
invariably 1 nt in Bacteria (always G) but 3 or 4 in Archaea. In addition, in
bacterial RNAs lacking P13/14, P12 lacks an otherwise highly conserved AA bulge
2 bp from the base of the helix. The euryarchaeal RNAs, which also lack P13/14,
are also interrupted near the base of the helix.
RNase P enzymes from all organisms share common ancestry; fundamental features of the structure and biochemistry should therefore be
preserved in all of its modern forms. Because the RNA components of archaeal
RNase Ps are not capable of catalysis and yet are sufficiently similar to those
of Bacteria for the identification of homologous sequences and higher order structures, they provide an important avenue for the investigation of essential
functions and structures in this RNA enzyme.
We thank N.R.Pace for helpful discussion during these investigations. This work
was supported by NIH grant 1-R29- GM52894-01 to JWB and NIH grant R01-GM48665-03 to CJD. BMV was supported by a North Carolina
Agriculture Foundation graduate fellowship. CJD is an associate of the Canadian
Institutes for Advanced Research.
REFERENCES
Return